Built on extensive hands-on experience, real-world consulting engagements, and deep technical research into attacking and defending AI systems. No theoretical fluff: learn precisely how AI systems can be compromised, how to assess their vulnerabilities methodically, and how to strengthen defenses against emerging threats.
If you're an offensive security professional, defender, or a technical leader looking to master the cutting edge of AI cybersecurity, this course is your next step.
$2,000 USDSep 22 & 24, 20262 live sessionsIntermediate Cyber, Basic AI
// level: intermediate cyber, basic AI · recorded & distributed to students · bulk purchases available
What you'll learn
Skills you'll walk away able to use.
Threat model enterprise AI deployments across LLM and image-based systems
Craft prompt injection attacks against public and custom LLM endpoints
Distinguish fuzzing (gradient-based) from logical prompt injection
Jailbreak production LLMs using documented bypass methods
Exploit AI-integrated applications and LLM-powered APIs
Attack RAG systems and autonomous AI agents
Map findings to MITRE ATLAS and the OWASP LLM Top 10
Assess AI supply-chain risk across third-party tools and open-source models
Run AI red team engagements using current industry methodologies
Apply the Arcanum LLM Assessment Methodology end to end
Build defense bypasses with the Arcanum Prompt Injection Taxonomy
Deploy defensive countermeasures and AI hardening techniques
What your employer gets
A team member who can assess AI systems for prompt injection, jailbreaks, and agent attacks
A repeatable methodology for reviewing LLM-integrated applications, the Arcanum LLM Assessment Methodology
Coverage mapped to MITRE ATLAS and the OWASP LLM Top 10 for consistent reporting
The Arcanum Prompt Injection Taxonomy for testing enterprise LLM defenses
Course recordings the attendee can revisit and share knowledge from across the team
Hands-On Labs
The most labs of any AI course.
Attacking AI isn't slideware. You'll attack a range of deliberately vulnerable, LLM-powered applications modeled on real products: e-commerce assistants, enterprise copilots, healthcare bots, coding agents, MCP gateways, and more. Many labs come with guided walkthroughs, so you can take them home and work through them at your own pace after class.
21
Proprietary Arcanum labs, built in-house
71
Curated open-source labs
92
Total hands-on labs, the most of any course in the industry
Guided
Walkthroughs for many of the labs, to take home or do after class
Amazoom.com
// e-commerce · AI shopping assistant
Stark HR Assistant
// enterprise copilot · RAG exfiltration
Instaglam
// social app · AI support bot
// modeled on real products · yours to keep practicing after class
Course Details
What's included.
Format
// delivery
Hybrid of lectures, interactive discussions, and hands-on labs
Live training and Q&A
Class recordings available online
Certificate of completion
Sep 22 & 24, 10AM to 5PM MST
Prerequisites
// level
Intermediate cybersecurity knowledge
Basic understanding of AI concepts
Recommended level: Intermediate Cyber, Basic AI
Community
// access
Private Discord channel
Channels shared with the Red Blue Purple AI course
Discussion and resource sharing
Who It's For
Built for practitioners at the cutting edge.
01
Offensive Security Professionals
Learn how AI systems can be compromised and add prompt injection, jailbreaking, and AI red teaming to your toolkit.
02
Defenders
Understand the attacks methodically so you can harden AI systems and deploy defensive countermeasures that hold up.
03
Technical Leaders
Master the risks of AI adoption across your stack and bring a structured assessment methodology back to your org.
Syllabus
Thirteen modules, hands-on throughout.
Note: this syllabus is subject to updates, as AI security is a rapidly evolving field.
M01The AI Gold Rush
Understanding rapid AI adoption and its cybersecurity implications
Exploring key industries integrating AI (finance, healthcare, gaming, automotive) and their unique risks
Analysis of traditional security vulnerabilities prevalent in AI-driven applications (e.g., input validation, authentication, authorization)
M02Common AI Architectures and Ecosystem Risks
Deep dive into the AI development pipeline: model selection criteria, training procedures, deployment strategies
Step-by-step guidance for crafting effective prompt injection scenarios
Exercises targeting various public and custom LLM endpoints
Discussion on practical defense mechanisms against prompt injection
M05LLM Jailbreaking for Security Professionals
Overview of LLM jailbreak methods
Review of notable jailbreak cases (ChatGPT, Claude, Gemini)
Practical considerations and implications of jailbreak attacks
M06Privacy and Ethical Considerations
Ethical hacking boundaries specific to AI
Privacy implications and compliance considerations (e.g., GDPR)
Responsible disclosure practices for AI vulnerabilities
M07AI Red Teaming Methodologies
Examination of current industry approaches and best practices in AI red teaming
Identification of key organizations and leaders driving AI security testing advancements
Case studies showcasing red teaming scenarios in diverse AI ecosystems
M08Attacking AI-Integrated Applications
Vulnerabilities and risks associated with AI-powered APIs
Detailed exploration of API security considerations specific to LLM-integrated systems
Real-world scenarios illustrating successful exploits of AI-integrated applications
M09MITRE ATLAS & OWASP AI Top Ten
Walkthrough of MITRE's ATLAS framework tailored for AI adversarial attacks
Breakdown of OWASP's Top 10 security vulnerabilities specific to LLM-based applications
M10Emerging Attack Techniques and Research
Overview of cutting-edge academic and industry research in AI security
Techniques for developing and innovating AI testing methodologies
Resources for continuous learning: key academic papers, blogs, and repositories
M11Defensive Countermeasures and AI Hardening
Strategies for strengthening AI systems against attacks
Defensive tooling specifically designed for AI environments
M12The Arcanum LLM Assessment Methodology
Introduction and step-by-step guide to Arcanum's structured methodology for assessing AI security
Details, best practices, and actionable guidelines for AI penetration testers
M13The Arcanum Prompt Injection Taxonomy
Finally we cover modern defenses in enterprise-based deployments of LLM enabled applications and our taxonomy to help testers devise bypasses to modern defenses. This includes going over our proprietary intent, technique, evasion, and utility format.
FINClosing Discussion & Practical Review
The course will be recorded and distributed to students after completion, so you can revisit the material on your own schedule.
Sneak Peek
Watch the talk version.
Want a taste before you enroll? Here's Jason's conference talk on attacking AI, a preview of the ideas this course takes hands-on across two live days.
Your Instructor
Taught by Jason Haddix.
JH
Jason Haddix
CEO · Lead Instructor
"Over the past few years, I've immersed myself in the intersection of offensive security and artificial intelligence, turning curiosity into actionable insights and practical methodologies. This journey has evolved into talks, research papers, and now this course. At Arcanum, our mission is to make a tangible impact on the security community with world class, modern, and accessible training."
Good to Know
Policies & access.
Refund & Access Policy
Because our training includes proprietary, cutting-edge content, all registrations are non-refundable. If you are unable to attend live sessions, full recordings will be provided following the conclusion of the course so you can access the material on your own schedule.
Class Collaboration
Participants will have access to private Discord channels shared with the "Red Blue Purple AI" course for discussion and resource sharing.
Bulk Purchases
Yes. Bulk purchases and team discounts are available. Reach out and we'll set up seats and bulk pricing for your organization.
 Attacking AI.c482709f.png)
